A low-carbon demand response resource allocation method based on a dynamic carbon emission factor

Abstract

The application discloses a kind of low-carbon demand response resource allocation methods based on dynamic carbon emission factor, steps are as follows:1, dynamic carbon emission factor is obtained by solving;2, based on Dirichlet process mixture model, dynamic carbon emission factor data and power consumption data are adaptively clustered, and the clustering label and its probability distribution of carbon emission and power consumption in different time periods are obtained;3, based on historical data and the result of 2, the power consumption and dynamic carbon emission factor of future 24 hours are predicted using double-LSTM;4, based on the fixed daily power consumption constraint and the calculation result of 3, an optimization model for carbon emission reduction demand response is constructed;5, based on rigid constraint and flexible constraint condition, the power consumption demand and carbon emission of future 24 hours are obtained by solving using demand response optimization model, which is used for configuring power consumption scheme.The application can not only accurately evaluate the contribution of demand response measures to carbon emission reduction, but also realize the collaborative optimization of user side and grid side, maximize the carbon emission reduction benefit.

Description

TECHNICAL FIELD

[0001] The present application relates to the technical field of low-carbon dispatching of power systems, and particularly relates to a low-carbon demand response resource optimization configuration method based on a dynamic carbon emission factor. BACKGROUND

[0002] With the increasingly serious global climate change problem, as one of the main sources of carbon emissions, the power system has higher requirements for its low-carbon transformation. The traditional demand response mechanism mainly focuses on the economy and reliability of the power system, and adjusts the power consumption behavior of users through incentives to smooth the load fluctuation and reduce the system operation cost, but fails to take carbon emissions as a core consideration factor. At the same time, the existing carbon emission evaluation mainly adopts a static carbon emission factor, which assumes that the carbon emission intensity of the generator set remains constant at different times, and ignores the influence of renewable energy output fluctuation, load dynamic change and power grid operation state on real-time carbon emission, resulting in a significant deviation between the carbon emission evaluation result and the actual situation.

[0003] At present, the power system is faced with the following main problems: first, the traditional demand response mechanism overemphasizes economic benefits, and only uses price signals to guide user behavior, without fully considering the carbon emission characteristics under different space-time conditions, so it is difficult to effectively support the low-carbon transformation of the power system; second, the static carbon emission factor calculation method cannot accurately reflect the real-time operation state of the power system, especially under the background of large-scale renewable energy grid connection, the fluctuation and intermittency characteristics of the power generation side and the dynamic change of the load side make the carbon emission present obvious spatial and temporal difference; third, there is a lack of resource optimization configuration method that deeply integrates dynamic carbon emission factor and demand response, and the existing dispatching model often takes economy as the main optimization target, without establishing an effective coupling mechanism between carbon emission and power dispatching, so that the low-carbon dispatching effect cannot meet the requirements of low-carbon emission.

[0004] Therefore, it has important theoretical value and practical significance to develop a low-carbon demand response resource optimization configuration technology based on a dynamic carbon emission factor. SUMMARY

[0005] The purpose of the present application is to provide a low-carbon demand response resource configuration method based on a dynamic carbon emission factor, which realizes the optimal timing scheduling of the power consumption load under the constraint of fixed power consumption by constructing a dynamic carbon emission factor model considering the time dimension, so as to minimize the overall carbon emission of the system.

[0006] To achieve the above purpose, the present application provides the following technical scheme:

[0007] On the one hand, the present application provides a low-carbon demand response resource configuration method based on a dynamic carbon emission factor, which includes the following steps:

[0008] S1, integrating the marginal carbon emission factor and the average carbon emission factor to obtain a dynamic carbon emission factor;

[0009] S2, adaptively clustering the dynamic carbon emission factor data and the electricity consumption data based on a Dirichlet process mixture model to obtain clustering labels and probability distributions of carbon emission and electricity consumption in different time periods, and intelligently identifying and classifying the carbon emission characteristics and electricity consumption characteristics in different time periods;

[0010] S3, predicting the electricity consumption and the dynamic carbon emission factor in the next 24 hours based on historical data and the result of S2 using a double-LSTM deep learning network;

[0011] S4, constructing a demand response optimization model with the maximum carbon emission reduction as the target based on the calculation result of S3 under the constraint of fixed daily electricity consumption;

[0012] S5, obtaining the electricity demand and the carbon emission in the next 24 hours based on the demand response optimization model under the conditions of rigid constraints and flexible constraints, and using the result to configure the electricity consumption scheme.

[0013] In some embodiments, the marginal carbon emission factor is:

[0014]

[0015] wherein,

[0016]

[0017] The average carbon emission factor AEF t reflects the average carbon emission level of the entire power grid at a specific time point, and the average carbon emission factor is:

[0018]

[0019] wherein,

[0020]

[0021] The is a piecewise function used to determine the capacity utilization rate of the power plant.

[0022] When , it means that when the cumulative installed capacity is less than the reserved capacity, the power plant is running at full load.

[0023] When , it means that when the cumulative installed capacity before has reached or exceeded the reserved capacity, the power plant stops running.

[0024] In other cases, it means partial load operation, running according to the proportion of the remaining demand to the installed capacity.

[0025] where εp,f is the unit carbon emission of the specific fuel of the pth power plant, p = 1, 2,..., P; ηp,f is the carbon emission factor of the specific fuel of the pth power plant; and ηp is the average carbon emission factor of the pth power plant. p T is the transmission efficiency; is the residual electricity load at time t; is the installed capacity of the ith generator set; is the installed capacity of the pth power plant; is the capacity utilization rate; is the electricity generated by each fuel type f at time step t, and F is the set of all power generation fuel types.

[0026] In some embodiments, the dynamic carbon emission factor is:

[0027] CEF t = α AEF t + (1 - α) MET t ;

[0028] where α is a weight coefficient, and α ∈ (0, 1).

[0029] In some embodiments, S2 comprises the following steps:

[0030] S21, constructing a Dirichlet process mixture model;

[0031] S22, training the Dirichlet process mixture model by a Gibbs sampling method until the model converges;

[0032] S23, obtaining the clustering labels and their probability distributions of each time period based on the trained Dirichlet process mixture model;

[0033] S24, identifying time periods with similar carbon emission characteristics according to the clustering results of S23.

[0034] In some embodiments, the Dirichlet process mixture model is:

[0035]

[0036] The carbon emission factor data generation process is:

[0037]

[0038] The electricity consumption data generation process is:

[0039]

[0040] where G is a random probability measure; DP(α, G0) is a Dirichlet process, α is a concentration parameter, and G0 is a base measure; θ i is the θth​i a parameter of observation data; x i is a random variable representing observation data; F(θ i ) is a likelihood function of observation data; π i is a mixture function satisfying N(μ i ,∑ i ) is a Gaussian distribution, μ i ,∑ i are the mean and covariance matrix of the i-th component, respectively; ω j is a mixing weight; v j ,Λ j are the mean and covariance matrix of the j-th component, respectively.

[0041] In some embodiments, in S3, after obtaining the feature categories of each time point data in the previous 7 days, the feature category information is fused with the original data when constructing the input of the LSTM optic nerve network prediction model. The input feature vector of the LSTM deep learning network for predicting the dynamic carbon emission factor is:

[0042] X cf (f) = [cf t , class cf (t)];

[0043] wherein class cf (t) = argmin i∈[1,k] d cf (t, i);

[0044] The input feature vector of the LSTM deep learning network for predicting the electricity consumption is:

[0045] X ec (f) = [ec t , class ec (t)];

[0046] wherein class ec (t) = argmin i∈[1,m] d ec (t, i);

[0047] In the formula, cf t is the dynamic carbon emission factor data at time t; class cf (t) is the feature category to which the dynamic carbon emission factor data at time t belongs; ec t is the electricity consumption data at time t; class ec (t) is the feature category to which the electricity consumption data at time t belongs; k represents the total number of feature categories of the dynamic carbon emission factor data; and m represents the total number of feature categories of the electricity consumption data.

[0048] In some embodiments, in S4, the demand response optimization model aiming at maximizing carbon emission reduction is:

[0049]

[0050] wherein CEF t is a predicted 24-hour dynamic carbon emission factor; is an electricity increase amount; is an electricity decrease amount, The electricity increase amount is equal to the electricity decrease amount, i.e., the electricity consumption is fixed.

[0051] In some embodiments, in S5, the rigid constraints include: daily electricity consumption conservation constraints, power limit constraints, device safe operation constraints; the flexible constraints include load change rate constraints, user comfort constraints.

[0052] In some embodiments, in order to ensure that the total electricity consumption remains unchanged before and after the demand response adjustment, the daily electricity consumption conservation constraints are:

[0053]

[0054] Considering the carrying capacity and safe operation requirements of the power system, the actual load in each time period must be controlled within the system-allowed range, and the power limit constraints are:

[0055]

[0056] In order to protect the safe operation of the electricity consumption device, the single load adjustment amplitude needs to be limited, and the device safe operation constraints are:

[0057]

[0058] In order to avoid the impact of severe load fluctuation on the power grid, the load change rate between adjacent time periods needs to be limited, and the load change rate constraints are:

[0059]

[0060] Considering the user experience, the influence of load adjustment on the daily life and production activities of users is controlled, and the user comfort constraints are:

[0061]

[0062] wherein, is an electricity increase amount; is an electricity decrease amount; is a reference load; P max , P minrespectively are the maximum and minimum allowed loads; is the maximum allowed power increase; is the maximum allowed power decrease;R r represents the maximum allowed difference of power change between adjacent time periods; and β is the user-acceptable load adjustment proportionality coefficient.

[0063] In another aspect, the present application provides a low-carbon demand response resource allocation system based on dynamic carbon emission factors, which uses the above method and includes the following modules:

[0064] A dynamic carbon emission factor calculation module: the dynamic carbon emission factor is calculated using the marginal carbon emission factor and the average carbon emission factor;

[0065] A clustering module: based on the Dirichlet process mixture model, the dynamic carbon emission factor data and the electricity consumption data are adaptively clustered to obtain the clustering labels and probability distributions of carbon emission and electricity consumption in different time periods;

[0066] A prediction module: based on the historical data and the results of S2, the electricity consumption and the dynamic carbon emission factor in the next 24 hours are predicted using a double-LSTM deep learning network;

[0067] A demand response optimization module: based on the fixed daily electricity consumption constraint, a carbon emission reduction demand response optimization model is constructed based on the calculation results of S3;

[0068] A demand response scheduling module: based on the rigid constraints and flexible constraints, the demand response optimization model is used to obtain the electricity demand and carbon emission in the next 24 hours, which is used to configure the electricity consumption scheme.

[0069] Compared with the prior art, the present application has the following beneficial effects:

[0070] The demand response method in the prior art mainly focuses on electricity price and electricity consumption, ignoring the dynamic characteristics of carbon emission. The present application introduces the dynamic carbon emission factor into the demand response optimization, so that the user can flexibly adjust the electricity consumption behavior according to the real-time carbon emission intensity of the power grid, and realize accurate carbon emission reduction.

[0071] In order to solve the problem of inaccurate identification of carbon emission characteristics and electricity consumption characteristics of the power system, the present application uses an adaptive clustering method of Dirichlet process mixture model to accurately capture the carbon emission and electricity consumption patterns in different time periods, and provides more reliable data support for demand response decision-making.

[0072] The present application uses a double-LSTM deep learning network to realize 24-hour joint prediction of electricity consumption and dynamic carbon emission factor, which significantly improves the prediction accuracy and provides more accurate input for demand response optimization.

[0073] In order to solve the problem that user experience and environmental protection benefits are difficult to balance in demand response optimization, the complete constraint system including user comfort, equipment safety and the like is established, so that the actual use demand of the user is guaranteed while the maximum carbon emission reduction is realized. BRIEF DESCRIPTION OF DRAWINGS

[0074] Figure 1 It is a whole flowchart of the present application;

[0075] Figure 2 It is a comparison diagram of total carbon emission before and after optimization in embodiment 1 of the present application;

[0076] Figure 3 It is a comparison diagram of 24-hour carbon emission before and after optimization in embodiment 1 of the present application;

[0077] Figure 4 It is a diagram of 24-hour load increase and decrease variation in embodiment 1 of the present application. DETAILED DESCRIPTION

[0078] The technical solutions in the embodiments of the present application will be described clearly and completely below with reference to the drawings in the embodiments of the present application. Obviously, the described embodiments are only part of the embodiments of the present application, rather than all the embodiments of the present application. Based on the embodiments in the present application, all other embodiments obtained by those skilled in the art without creative labor fall within the protection scope of the present application.

[0079] Embodiment 1:

[0080] Please refer to Figure 1 A low-carbon demand response resource allocation method based on a dynamic carbon emission factor, by constructing a dynamic carbon emission factor evaluation model considering the time and space characteristics, accurately depicting the carbon emission characteristics of the power system under different operating states, and introducing the carbon emission index as the core element into the demand response optimization decision. Not only can the contribution of demand response measures to carbon emission reduction be accurately evaluated, but also the user side and the grid side can be optimized cooperatively, while the economic and reliable performance of the system is ensured, and the carbon emission reduction benefit is maximized. At the same time, the present embodiment provides a new technical path for low-carbon transformation of the power system, which helps to improve the low-carbon operation level of the power system and promote the deep participation of the demand side resources in system regulation.

[0081] The specific steps are as follows:

[0082] S1, the dynamic carbon emission factor is integrated by using the marginal carbon emission factor and the average carbon emission factor.

[0083] The multi-source data of power system is searched and integrated, including the generator unit characteristics, fuel emission amount, power consumption amount of a certain area in a certain year, etc. The piecewise linear method is adopted to perform piecewise linear fitting on the unit generation efficiency curve and transmission efficiency curve to obtain the generation efficiency and transmission efficiency, and the dynamic marginal carbon emission factor and average carbon emission factor are calculated from the available data.

[0084] The unit carbon emission amount of a certain fuel of a certain power plant is given by the following formula:

[0085]

[0086] In the formula, ε f is the carbon emission intensity of a certain fuel type f; is the generation efficiency of the power plant p for this fuel.

[0087] The marginal carbon emission factor MET t over time is:

[0088]

[0089] wherein,

[0090]

[0091] In the formula, is the carbon emission amount of the power plant in a given time step.

[0092] The marginal carbon emission factor MET t reflects the incremental carbon emission when the load changes, and MET

[0093] The average carbon emission factor AEFtreflects the average carbon emission level of the entire power grid at a certain time point:

[0094]

[0095] wherein,

[0096]

[0097] The is a piecewise function for determining the capacity utilization rate of the power plant.

[0098] When , it indicates that the power plant is full-load operation when the cumulative installed capacity is less than the reserved capacity.

[0099] When , it indicates that the power plant stops operation when the previous cumulative installed capacity has reached or exceeded the reserved capacity.

[0100] Other cases represent partial load operation, running according to the proportion of the remaining demand and installed capacity.

[0101] The calculation expression of the dynamic carbon emission factor is:

[0102] CEF t =αAEF t +(1-α)MET t ;

[0103] In the formula, α is a weight coefficient, α ∈ (0, 1), which is determined according to the degree of load change.

[0104] The dynamic carbon emission factor is calculated by weighted fusion of the marginal carbon emission factor and the average carbon emission factor, which has significant advantages. The embodiment can capture both the instantaneous dynamic characteristics and the long-term steady-state characteristics of the power system: MET t reflects the short-term carbon emission impact caused by load change, and is suitable for evaluating the emission reduction effect of short-term regulation measures such as demand response; AEF t represents the overall carbon emission level of the system, and is suitable for evaluating the effect of long-term emission reduction strategies such as clean energy replacement. By adjusting the weight coefficient α, the influence degree of short-term and long-term characteristics can be flexibly balanced according to the actual application scene. When the load fluctuates violently, the weight of MET t is increased to highlight the marginal effect, and when the system is relatively stable, the weight of AEF t is increased to reflect the overall characteristics. This weighted fusion mechanism provides a more comprehensive and accurate carbon emission evaluation framework, which can provide a more reliable theoretical basis for low-carbon dispatching decisions of the power system.

[0105] S2, based on a Dirichlet Process Mixture Model (DPMM), adaptively clusters the dynamic carbon emission factor data and the electricity consumption data to obtain clustering labels and probability distributions of carbon emission and electricity consumption in different time periods, and realizes intelligent identification and classification of carbon emission characteristics and electricity consumption characteristics in different time periods. The Dirichlet Process Mixture Model is a non-parametric Bayesian model.

[0106] In view of the complex nonlinear characteristics of the dynamic carbon emission factor and the electricity consumption data, a non-parametric Bayesian clustering method is used for modeling analysis.

[0107] Let the dynamic carbon emission factor data set calculated by S1 be: CEF={cef1,cef2,...,cef n};

[0108] The electricity consumption data is: EC={ec1,ec2,...,ec n}

[0109] Wherein n is the number of samples.

[0110] Because the carbon emission characteristics in the power system are influenced by multiple factors such as power structure and renewable energy output, and the power load shows strong randomness and volatility, the traditional clustering method is difficult to effectively mine the potential rules in the data. Therefore, a Dirichlet process mixture model DPMM is used for clustering analysis. The Dirichlet process mixture model DPMM is expressed as:

[0111]

[0112] The carbon emission factor data generation process is:

[0113]

[0114] The power consumption data generation process is:

[0115]

[0116] The posterior distribution of the Dirichlet process mixture model DPMM is inferred by Gibbs sampling:

[0117]

[0118] In the formula, θ i is; θ -i is all parameters except θ i ; f(x i | θ i ) is a likelihood function; δ(θ j ) is a Dirac function; x 1:n is all observation data sequences; x i is the i th observation data; α is a mixing weight parameter; G0 is a reference distribution.

[0119] Through the Dirichlet process mixture model DPMM, the clustering results of the carbon emission factor data CEF clusters ={CEF1, CEF2,..., CEF k} and the power consumption data EC clusters ={EC1, EC2,..., EC m} can be obtained, wherein k and m are the best clustering numbers adaptively determined.

[0120] This embodiment fully considers the dynamics and uncertainty of carbon emissions in power systems, and realizes the adaptive identification of carbon emission patterns through a non-parametric Bayesian method, providing a new analysis tool for building demand response mechanisms based on dynamic carbon emission factors. The Dirichlet process mixture model (DPMM) can effectively capture and reflect the influence of multi-dimensional time and social factors on carbon emissions when performing clustering analysis on dynamic carbon emission factors. In the time dimension, it can identify the differences in carbon emission patterns between weekdays and holidays, as well as the changes in electricity load caused by seasonal changes. In the intra-day time sequence, it can distinguish the emission characteristics of typical periods such as morning and evening peaks, and industrial production-intensive periods. At the same time, it can also reflect the fluctuations in carbon emissions caused by major social activities (such as important events and large-scale exhibitions), extreme weather events, and sudden public events. In addition, the adaptive clustering feature of the Dirichlet process mixture model (DPMM) enables it to capture the influence of medium and long-term social development trends such as regional industrial structure adjustment, changes in energy consumption habits, and the popularization of new energy vehicles on carbon emission patterns, providing data support for developing more accurate demand response strategies. The results can be used to guide users to flexibly adjust their electricity consumption behavior based on the carbon emission characteristics of different time periods, promoting the low-carbon operation of the power system.

[0121] S3, based on historical data and the results of S2, use a double-LSTM deep learning network to predict electricity consumption and dynamic carbon emission factors for the next 24 hours.

[0122] When constructing the input features of the LSTM deep learning network prediction model, it is first necessary to determine the feature categories of the previous 7 days of historical data. These historical data need to be calculated separately with the feature class centers obtained by the Dirichlet process mixture model (DPMM) clustering to determine their respective feature categories.

[0123] For dynamic carbon emission factor data, assuming that the Dirichlet process mixture model (DPMM) clustering obtains k feature class centers C = {C1, C2,..., C k k}, the distance between the carbon emission factor data at time t and the i-th class center can be represented as:

[0124] d cf (t,i)=‖cf t -c i ‖;

[0125] For the carbon emission factor data at time t, its feature category can be determined by the following method:

[0126] class cf (t)=argmin i∈[1,k] d cf (t,i);

[0127] Similarly, for the electricity consumption data, assuming that the Dirichlet process mixture model DPMM clustering obtains m feature class centers, the distance between the electricity consumption data of each time t and the jth class center can be expressed as:

[0128] d ec (t,j)=||ec t -e j ||;

[0129] The corresponding feature class is determined as:

[0130] class ec (t)=argmin i∈[1,m] d ec (t,i);

[0131] Through the above calculation, the feature class of each time data in the previous 7 days can be obtained. When constructing the input of the LSTM deep learning network prediction model, these feature class information is fused with the original data. For the LSTM model used for dynamic carbon emission factor data prediction, the input feature vector can be expressed as:

[0132] X cf (f)=[cf t ,class cf (t)];

[0133] The input feature vector of the LSTM model for electricity consumption prediction is:

[0134] X ec (f)=[ec t ,class ec (t)];

[0135] In the design of the LSTM deep learning network structure, the LSTM deep learning network for electricity consumption prediction and carbon emission factor prediction is constructed respectively. Each network adopts a three-layer LSTM architecture: the first layer is configured with 128 LSTM units, mainly responsible for extracting basic features from complex input sequences; the second layer is set with 64 units, which optimizes the features extracted by the first layer; the third layer contains 32 units, which performs deep feature extraction and abstraction. Finally, through the full connection layer, the electricity consumption prediction value and the carbon emission factor prediction value of the next 24 hours are output respectively. In the training process, all input data are standardized to convert different dimensional features into the same scale range, so as to improve the training effect and prediction accuracy of the model.

[0136] The establishment of a demand response optimization model using predicted data for the next 24 hours can help users plan and adjust their electricity consumption behavior in advance. Through predictive power load management, users can shift high-power consumption activities from peak to off-peak periods, avoiding high electricity prices and reducing carbon emissions.

[0137] S4, based on the fixed daily electricity consumption constraint, the calculation results of S3 are used to build a carbon emission reduction demand response optimization model.

[0138] Assuming that the electricity consumption in the next 24 hours is fixed, a demand response optimization model is established based on the data predicted by S3, with the goal of maximizing carbon emission reduction. The objective function is:

[0139]

[0140] Where, The increase in electricity consumption is equal to the decrease in electricity consumption, i.e., the electricity consumption is fixed.

[0141] By intelligently adjusting the electricity consumption period, carbon emission reduction is achieved. This approach not only ensures user experience but also fully utilizes the difference in carbon emission intensity of power generation at different times, maximizing emission reduction while maintaining total electricity consumption.

[0142] S5, based on rigid and flexible constraints, the demand response optimization model is used to solve the electricity demand and carbon emissions in the next 24 hours, which is used to configure the electricity consumption scheme.

[0143] In the process of low-carbon demand response resource optimization configuration, the construction of constraint system is the key to ensure the feasibility of the scheme. These constraints can be divided into two categories: rigid constraints and flexible constraints.

[0144] Rigid constraints include the following constraints:

[0145] (1) Daily electricity conservation constraint:

[0146] Ensure that the total electricity consumption remains unchanged before and after demand response adjustment.

[0147]

[0148] (2) Power limit constraint:

[0149] Considering the carrying capacity and safe operation requirements of the power system, the actual load in each time period must be controlled within the system's allowable range:

[0150]

[0151] The power limit constraint prevents system overload or unstable operation caused by load adjustment.

[0152] (3) Equipment safety operation constraints:

[0153] To protect the safe operation of electrical equipment, the single load adjustment range needs to be limited.

[0154]

[0155] Existing methods often ignore the equipment bearing capacity, which can easily cause equipment damage. This embodiment considers the equipment safety constraints to extend the service life of the equipment and reduce the risk of equipment failure.

[0156] Flexible constraints include the following constraints:

[0157] (1) Load change rate constraint:

[0158] In order to avoid the impact of severe load fluctuations on the power grid, the load change rate between adjacent time periods needs to be limited:

[0159]

[0160] The load change rate constraint helps maintain the smooth operation of the system and reduces the power grid fluctuations caused by rapid load changes. r The setting needs to consider the system regulation capacity and stability requirements.

[0161] (2) User comfort constraint:

[0162] This constraint mainly considers user experience and controls the impact of load adjustment on user daily life and production activities:

[0163]

[0164] Where β is the user-acceptable load adjustment proportion coefficient.

[0165] The user comfort constraint ensures that the load adjustment is within the user-acceptable range and avoids significant interference with normal power consumption. Different β values can be set according to the characteristics of different types of users (such as residential, commercial, and industrial users).

[0166] By reasonably setting and dynamically adjusting these constraints, the optimal low-carbon demand response scheduling scheme can be achieved while ensuring system safety, user comfort, and equipment reliability.

[0167] S4 and S5 together constitute the optimization solving algorithm, and the optimal low-carbon demand response resource scheduling based on dynamic carbon emission factors is finally solved.

[0168] For example, Figure 2As shown in FIG. 4, first, the future 24-hour electricity load data and grid carbon emission factor data are obtained by a prediction algorithm. The predicted hourly electricity consumption is multiplied by the carbon emission factor of the corresponding period to obtain the baseline carbon emission before optimization. Subsequently, the 24-hour electricity load is redistributed by an optimization scheduling algorithm, and the optimized electricity scheme is obtained under the premise of ensuring electricity demand.

[0169] Figure 2 The bar chart in FIG. 5 shows the change in total carbon emissions over 24 hours before and after optimization scheduling. The total carbon emissions of the system before optimization are about 1.33 units, and after load optimization scheduling, the total carbon emissions are reduced to about 1.28 units, achieving a reduction of about 3.8%. Figure 3 The detailed comparison of hourly carbon emissions before and after optimization is shown in FIG. 6, which can intuitively show the emission reduction effect in different periods. Figure 4 The change analysis chart in FIG. 7 shows the specific increase and decrease of carbon emissions in each period within 24 hours, helping to analyze the effect distribution of optimization scheduling. Through this optimization scheduling scheme based on prediction data, the electricity load can be arranged in periods with lower grid carbon emission factors, thereby reducing the overall carbon emissions.

[0170] Embodiment 2

[0171] A low-carbon demand response resource allocation system based on dynamic carbon emission factors uses the above method and includes the following modules:

[0172] Dynamic carbon emission factor calculation module: the dynamic carbon emission factor is calculated using the marginal carbon emission factor and the average carbon emission factor;

[0173] Clustering module: based on the Dirichlet process mixture model, the dynamic carbon emission factor data and electricity consumption data are adaptively clustered to obtain the clustering labels and probability distribution of carbon emission and electricity consumption in different periods;

[0174] Prediction module: based on historical data and the results of S2, the electricity consumption and dynamic carbon emission factor in the next 24 hours are predicted using a double-LSTM deep learning network;

[0175] Demand response optimization module: based on the fixed daily electricity consumption constraint, a carbon emission reduction demand response optimization model is constructed based on the calculation results of S3;

[0176] Demand response scheduling module: based on the rigid constraint and flexible constraint conditions, the demand response optimization model is used to solve the electricity demand and carbon emissions in the next 24 hours to configure the electricity scheme.

[0177] The low-carbon demand response resource allocation system based on a dynamic carbon emission factor can be installed in a computer device. The computer device comprises a processor, a memory, and a computer program stored in the memory and executable on the processor, such as a low-carbon demand response resource allocation program based on a dynamic carbon emission factor. The memory comprises at least one type of readable storage medium, such as a flash memory, a mobile hard disk, a multimedia card, a card-type memory (such as an SD or DX memory), a magnetic memory, a magnetic disk, an optical disk, etc. The processor is the control core of the electronic device, connects various components of the computer device through various interfaces and lines, and executes various functions and processes data of the computer device by running or executing programs or modules stored in the memory and calling data stored in the memory.

[0178] The module refers to a series of computer program segments capable of being executed by the processor of the computer device and capable of completing a fixed function, which are stored in the memory of the computer device.

[0179] Other embodiments of the application will be apparent to those skilled in the art from consideration of the specification and practice of the application disclosed herein. The specification and examples given herein are intended as illustrative only and not as limitations on the present application.

Claims

1. A low-carbon demand response resource allocation method based on dynamic carbon emission factors, characterized in that, The method comprises the following steps: S1, integrating a marginal carbon emission factor and an average carbon emission factor to obtain a dynamic carbon emission factor; the marginal carbon emission factor is: ； wherein, ； the average carbon emission factor is: ； wherein, ； In the formula, For the first Power plant The carbon emissions per unit of a specific fuel, ; For transmission efficiency; for Residual electrical load at any given time; For the first The installed capacity of each generator set; For power plants The installed capacity; For capacity utilization; For each fuel type time step The amount of electricity generated, It is a set of all types of fuel used for power generation; the dynamic carbon emission factor is: ； In the formula, is a weight coefficient, ; S2, adaptively clustering the dynamic carbon emission factor data and the electricity consumption data based on a Dirichlet process mixture model to obtain clustering labels and probability distributions of carbon emission and electricity consumption in different time periods; S3, predicting the electricity consumption and the dynamic carbon emission factor in the next 24 hours based on historical data and the result of S2 by using a double-LSTM deep learning network; the input feature vector of the double-LSTM deep learning network is: ； wherein ; the input feature vector of the LSTM deep learning network for predicting the electricity consumption is: ； wherein ; In the formula, is dynamic carbon emission factor data of the moment; is a feature category to which the dynamic carbon emission factor data of the moment belongs; is power consumption data of the moment; is a feature category to which the power consumption data of the moment belongs; denotes the total number of feature categories of the dynamic carbon emission factor data; denotes the total number of feature categories of the power consumption data; is the distance between the carbon emission factor data of each moment and the first class center; is the distance between the power consumption data of each moment and the first class center; S4, constructing an optimization model of demand response with the maximum carbon emission reduction as the target based on the fixed daily electricity consumption constraint and the calculation result of S3; S5, obtaining the electricity demand and the carbon emission in the next 24 hours by solving the demand response optimization model based on the rigid constraint and the flexible constraint condition, and using the result to configure an electricity consumption scheme.

2. The low-carbon demand response resource allocation method based on dynamic carbon emission factors according to claim 1, characterized in that, S2 comprises the following steps: S21, constructing a Dirichlet process mixture model; S22, training the Dirichlet process mixture model by the Gibbs sampling method until the model converges; S23, obtaining the clustering labels and the probability distributions of each time period based on the trained Dirichlet process mixture model; S24, identifying the time periods with similar carbon emission characteristics according to the clustering result of S23.

3. The low-carbon demand response resource allocation method based on dynamic carbon emission factors according to claim 2, characterized in that, The Dirichlet process mixture model is: ； The carbon emission factor data generation process is: ； The electricity consumption data generation process is: ； In the formula, It is a measure of random probability; For the Dirichlet process, For concentration parameters, As a benchmark measure; For the first Parameters of each observation data; It is a random variable representing the observed data; Let be the likelihood function of the observed data; To meet The mixture function; It follows a Gaussian distribution. The first The mean and covariance matrix of each component; Mixed weights; The first The mean and covariance matrix of each component.

4. The low-carbon demand response resource allocation method based on dynamic carbon emission factors according to claim 1, characterized in that, In S4, the demand response optimization model with the maximum carbon emission reduction as the target is: ； In the formula, is a dynamic carbon emission factor; is an electricity consumption increase amount; is an electricity consumption decrease amount, The electricity consumption increase amount is equal to the electricity consumption decrease amount, i.e., the electricity consumption is fixed.

5. The low-carbon demand response resource allocation method based on dynamic carbon emission factors according to claim 1, characterized in that, In S5, the rigid constraint comprises a daily electricity consumption conservation constraint, a power limit constraint and a device safe operation constraint; the flexible constraint comprises a load change rate constraint and a user comfort constraint.

6. The low-carbon demand response resource allocation method based on dynamic carbon emission factors according to claim 5, characterized in that, The daily electricity consumption conservation constraint is: ； The power limit constraint is: ； ； The device safe operation constraint is: ； ； The load change rate constraint is: ； The user comfort constraint is: ； wherein is the increase in electricity consumption; is the decrease in electricity consumption; is the base load; , are the maximum and minimum loads allowed, respectively; is the maximum allowed increase in power; is the maximum allowed decrease in power; denotes the maximum allowed difference in power change between adjacent time periods; is the user-acceptable load adjustment proportionality factor.

7. A low carbon demand response resource allocation system based on dynamic carbon emission factors, using the method of any one of claims 1-6, characterized in that, The method comprises the following modules: a dynamic carbon emission factor calculation module for calculating the dynamic carbon emission factor by using the marginal carbon emission factor and the average carbon emission factor; a clustering module for adaptively clustering the dynamic carbon emission factor data and the electricity consumption data based on the Dirichlet process mixture model to obtain the clustering labels and the probability distributions of carbon emission and electricity consumption in different time periods; a prediction module for predicting the electricity consumption and the dynamic carbon emission factor in the next 24 hours based on historical data and the result of S2 by using a double-LSTM deep learning network; a demand response optimization module for constructing a demand response optimization model with the maximum carbon emission reduction as the target based on the fixed daily electricity consumption constraint and the calculation result of S3; a demand response scheduling module for obtaining the electricity demand and the carbon emission in the next 24 hours by solving the demand response optimization model based on the rigid constraint and the flexible constraint condition, and using the result to configure an electricity consumption scheme.

Images